Blog Writer — 深度博客写作助手 · v1.0.1

针对特定主题撰写结构完整、逻辑清晰、深入浅出的博客文章。支持七类写作场景，每类有专属写作模板与规范。在遇到复杂概念时自动调用 Concept Decoder skill；在科研与论文类文章中强制引用文献；在编程类文章中包含最佳实践与反模式分析。

Overview

This skill produces well-structured, intellectually rigorous blog posts on a given topic. It combines breadth (historical context, cross-domain connections) with depth (precise definitions, mathematical derivations where applicable, critical analysis), and adapts its structure and standards to the type of article being written.

Language policy: Respond in the same language the user writes in. Mixed input defaults to Chinese.

When to Use This Skill

Use /blog to trigger this skill when:

• User wants to write or draft a blog post on a specific topic
• User wants a structured, publication-ready article (not a quick answer)
• User says things like "帮我写一篇关于X的博客"、"write a blog post about X"、"draft an article on X"

Do NOT use this skill when:

• User only needs a brief explanation (use direct answer or /decode instead)
• User needs a formal academic paper (different genre conventions apply)

Trigger Syntax

/blog [topic]
/blog [topic], [category]
/blog [topic], [category], [length]

Examples:

/blog Python异步编程
/blog 肿瘤微环境中的细胞通讯, 科学研究
/blog 贝叶斯定理, 数学概念, long
/blog Attention Is All You Need 论文解读
/blog 费曼学习法, 思维方法
/blog 为什么睡眠如此重要, 科普
/blog 《哥德尔、艾舍尔、巴赫》书评

If category is not specified, infer from the topic. If ambiguous, ask the user to confirm.

Article Length Levels

Universal Writing Principles

These apply to all article categories without exception.

P1 — Origin and History First · 溯源优先

Every article must include the origin and development history of the topic:

• Who first proposed it? When? In what context?
• How has it evolved? What were the key turning points?
• What is the current state of the field?
• Verify all factual claims: names, dates, institutions, events must be accurate.

P2 — Systemic, Global Perspective · 系统视角

Treat the topic as part of a larger system:

• What does this concept connect to? What does it depend on?
• What broader trends or frameworks does it belong to?
• Avoid treating the topic as an isolated island.

P3 — Mandatory Critical Thinking · 批判性思维（强制）

Every article must present both sides of the argument:

• What are the strengths, successes, and evidence in favor?
• What are the limitations, criticisms, failure cases, or open questions?
• Explicitly label contested claims as contested; do not present one view as universal truth.
• For scientific topics: distinguish established consensus from active debate.

P4 — Precise Definitions · 精确定义

For every important concept introduced:

• Provide a clear, accurate definition on first use
• If the concept is complex or counterintuitive → trigger Concept Decoder:

  [CALL: concept-decoder @ https://clawhub.ai/onlybelter/concept-decoder]
  /decode [concept name]
  

• Summarize key concepts in a Glossary section at the end (for Long articles)

P5 — Mathematics with Intuition · 数学 + 直觉

When mathematical formulas or derivations are included:

• Define every symbol before use
• Show the derivation incrementally (don't jump steps)
• Always follow a formula with a plain-language interpretation
• Use visual descriptions or analogies to build intuition
• Mark the conceptually critical step with ⚡
• Maximum formula density: no more than 1 display equation per 200 words on average

P6 — Fact Verification · 事实核查

Before finalizing any article, verify each item:

• 所有人名：全名、所属机构、国籍是否正确？· All named persons: full name, affiliation, nationality correct?
• 所有日期：发表/发明/事件年份是否正确？· All dates: year of publication/invention/event correct?
• 所有机构名：拼写是否正确，是否仍然存在？· All institutional names: spelled correctly, still active?
• 所有引用统计数据：来源是否可追溯？· All cited statistics: source traceable?
• 所有引用语录：是否经原始来源核实？· All attributed quotes: verified against original source?

Article Structure Template (Universal)

Every article follows this spine, with category-specific sections inserted at marked positions:

1. Hook / Opening
2. [CATEGORY-SPECIFIC: Context Block]
3. Historical Background & Development
4. Core Content
   ├── [CATEGORY-SPECIFIC sections]
   ├── Mathematical Content (if applicable) [P5]
   └── Critical Analysis [P3]
5. [CATEGORY-SPECIFIC: Special Section]
6. Summary & Key Takeaways
7. Further Reading / References
8. [OPTIONAL] Glossary

Category-Specific Templates

\x3C!-- [I-01 FIX] Category order aligned with README: 论文解读 → 科学研究 → 编程技术 → 数学概念 → 思维方法 → 书评 → 科普 -->

📄 Category 1: 科学论文解读 (Paper Walkthrough)

Trigger keywords: "论文解读"、"paper walkthrough"、"解读这篇论文"、"读懂X论文"

Goal: Help readers understand a specific scientific paper — its context, contributions, methods, results, and significance — without requiring them to read the full paper first.

Required Sections

§1 — Paper Identity Card

标题：
作者：（第一作者 + 通讯作者，其余可省略）
期刊/会议：
发表年份：
DOI / arXiv链接：
引用数（截至写作时）：
一句话概括：

§2 — Why This Paper Matters

• What problem was the field struggling with before this paper?
• Why was this paper a breakthrough (or why is it controversial)?
• Who should read this paper and why?

§3 — Background: What You Need to Know First

• List 3–5 prerequisite concepts
• For each complex prerequisite → trigger Concept Decoder
• Cite 2–3 key background papers

§4 — Paper Structure Roadmap

• Brief guide to the paper's sections (what each section does, not what it says)
• Highlight which sections are essential vs. skippable for different readers

§5 — Core Contributions (the "What")

• List the paper's main claims/contributions as numbered points
• For each: state it precisely, then explain it in plain language

§6 — Methods Deep Dive (the "How")

• Explain the key methodological innovations
• Include core equations with full symbol definitions and intuitive explanations [P5]
• Flag any methodological assumptions or limitations

§7 — Results and What They Mean

• Key experimental/theoretical results
• What do the numbers/figures actually tell us?
• Compare to prior state-of-the-art

§8 — Critical Evaluation [P3 — MANDATORY]

• ✅ Strengths: what the paper does well
• ⚠️ Limitations: what the paper does not address or assumes away
• ❓ Open questions: what remains unresolved after this paper
• 🔁 Subsequent work: has the community validated, extended, or challenged these results?

§9 — Impact and Legacy

• Citation trajectory and field impact
• Papers that directly built on this work (cite 2–3)
• Has the paper's influence grown or faded over time?

Citation standard: Cite the paper being decoded as [Paper] throughout; all other references as [Author, Year].

🔬 Category 2: 科学研究 (Scientific Research)

Trigger keywords: 科研综述、研究进展、领域综述、research overview

Goal: A rigorous, balanced overview of a research topic — suitable for researchers entering a field or experts wanting a structured synthesis.

Required Sections

§1 — The Central Question

• State the core scientific question this field is trying to answer
• Why does it matter? (scientific significance + broader implications)

§2 — Historical Development [P1 — MANDATORY]

• Timeline of key milestones (use a text-based timeline or table)
• Founding figures and their contributions (verify names/dates [P6])
• Paradigm shifts: what changed the field's direction?

§3 — Current State of Knowledge

• What is firmly established (consensus)?
• What is actively debated?
• What is unknown?
• Use a Knowledge Map table:

  | Status | Claims |
  |--------|--------|
  | ✅ Established consensus | ... |
  | 🔄 Active debate | ... |
  | ❓ Open question | ... |
  

§4 — Key Methods and Tools

• Dominant experimental / computational / theoretical approaches
• Include core equations where central to the field [P5]
• For complex methods → trigger Concept Decoder

§5 — Critical Analysis [P3 — MANDATORY]

• Major controversies in the field
• Reproducibility concerns (if any)
• Methodological blind spots
• Alternative interpretations of key results

§6 — Future Directions

• Most promising open problems
• Emerging methods or paradigm shifts on the horizon

Citation standard: MANDATORY. Cite recent papers (≤5 years preferred) AND seminal high-citation works. Format: [Author et al., Year, Journal]. Full list at end.

💻 Category 3: 编程技术 (Programming & Technology)

Trigger keywords: 编程、代码、框架、库、算法实现、技术选型

Goal: A technically accurate, practically useful article that helps developers understand and correctly apply a technology.

Required Sections

§1 — What Problem Does This Solve?

• The concrete pain point this technology addresses
• What existed before and why it was insufficient

§2 — Historical Context & Evolution [P1 — RECOMMENDED]

• Origin of the technology (creator, year, motivation)
• Major version milestones and what changed
• Current ecosystem status (actively maintained? deprecated? fragmented?)

§3 — Core Concepts

• Key abstractions and their precise definitions [P4]
• Mental model: how should a developer think about this?
• For complex concepts → trigger Concept Decoder

§4 — How It Works (Internals)

• Architecture / mechanism explanation
• Include diagrams (ASCII or described) where helpful
• Mathematical foundations if relevant [P5]

§5 — Code Examples

• Minimal working example first (the "hello world")
• Progressive complexity: basic → intermediate → advanced
• All code must be syntactically correct and runnable
• Annotate non-obvious lines

§6 — ✅ Best Practices [MANDATORY for all programming articles]

• Numbered list of recommended patterns
• For each: explain why it's a best practice, not just what to do
• Include performance, security, and maintainability considerations

§7 — ❌ Anti-Patterns: What to Avoid [MANDATORY for all programming articles]

• Common mistakes and misuses
• Each anti-pattern follows this fixed format:

\x3C!-- [I-04 FIX] Added concrete code-block format example, consistent with README -->

# ❌ Anti-pattern: [name]
# Problem: [what goes wrong and why]
[code showing the bad pattern]

# ✅ Fix:
[correct approach]

• Include subtle pitfalls that even experienced developers miss

§8 — Critical Evaluation [P3 — MANDATORY]

• When to use this technology (and when NOT to)
• Trade-offs vs. alternatives
• Known limitations, edge cases, version compatibility issues

Citation standard: Official documentation, RFC/PEP/spec documents, authoritative blog posts (e.g., engineering blogs from major tech companies). Format: [Source Name, URL, accessed date].

📐 Category 4: 数学概念 (Mathematical Concepts)

Trigger keywords: 数学、定理、证明、公式、代数、分析、几何、概率

Goal: Make a mathematical concept genuinely understandable — not just formally correct, but intuitively grasped.

Primary tool: This category has the highest overlap with Concept Decoder. For the central concept, always trigger Concept Decoder first, then expand into the full article structure.

Required Sections

§1 — The Problem This Math Solves [P1 — MANDATORY]

• What question or difficulty motivated this concept?
• What failed before it existed?

§2 — Historical Development [P1 — MANDATORY]

• Origin: who, when, in what context?
• Key contributors and their roles
• Surprising historical facts (e.g., simultaneous independent discovery, long delays between invention and application)

§3 — Intuitive Foundation

• Everyday analogy (concrete, visual)
• Where the analogy breaks — be explicit
• Cross-domain analogy (structural parallel in another field)

§4 — Formal Development [P5 — MANDATORY]

• Precise definition(s) — if multiple equivalent definitions exist, present all and explain why they're equivalent
• Key theorems with proof sketches (full proofs in appendix/footnote for Long articles)
• Incremental formula build-up with ⚡ marking the critical step
• Plain-language interpretation after every display equation

\x3C!-- [I-05 FIX] Added category-specific math formatting block, consistent with README --> Math Formatting Rules (this category):

• Display equations: $$...$$, centered; number them if cross-referenced
• Inline math: $...$
• Mark the conceptually critical derivation step with ⚡
• For articles with 5+ equations: include a Symbol Table before §4

§5 — Examples and Special Cases

• Simplest non-trivial example first
• Boundary cases: what happens at the limits?
• Numerical examples where illuminating (code optional)

§6 — Connections and Generalizations

• What broader framework contains this concept?
• What does it reduce to in special cases?
• Surprising appearances in other fields (lateral connections)

§7 — Critical Perspective [P3 — MANDATORY]

• Alternative definitions or formulations (and why they differ)
• Historical controversies (e.g., debates over rigor, constructivism vs. formalism)
• Where the concept breaks down or requires extension

Citation standard: Cite original papers AND standard textbooks. Format: Author, Title, Publisher/Journal, Year.

🧠 Category 5: 思维方法 (Mental Models & Thinking Methods)

Trigger keywords: 思维、方法论、学习、认知、决策、心智模型

Goal: Present a thinking framework in a way that is intellectually honest, practically actionable, and resistant to oversimplification.

Required Sections

§1 — The Cognitive Problem

• What failure mode in human thinking does this method address?
• Concrete example of the problem (a story or scenario)

§2 — Origin and Intellectual History [P1 — RECOMMENDED]

• Who developed this framework? In what field?
• Has it been validated empirically? (cognitive science, psychology, education research)
• For concepts with scientific backing → cite relevant research

§3 — The Framework Explained

• Clear, precise definition [P4]
• Step-by-step breakdown
• For complex cognitive concepts → trigger Concept Decoder

§4 — Worked Examples

• At least 2 concrete examples from different domains
• Show the method being applied, not just described

§5 — Evidence and Effectiveness

• What does the research say? (cite if available)
• Anecdotal vs. empirical evidence — distinguish clearly

§6 — Critical Analysis [P3 — MANDATORY]

• ✅ When this method works well
• ⚠️ When it fails or backfires
• 🚫 Common misapplications and oversimplifications
• Competing frameworks and how they compare

§7 — Practical Application

• Concrete, actionable steps to implement this method
• Common obstacles and how to overcome them
• How to know if it's working

Citation standard: Cognitive science and psychology papers where available. Popular books cited as secondary sources. Format: Author, Title, Year.

📚 Category 6: 书评 / 文献综述 (Book Review / Literature Survey)

Trigger keywords: 书评、读书笔记、文献综述、review、读后感

Goal: A critical, structured evaluation that helps readers decide whether to read the work and what to take from it.

Required Sections

§1 — Work Identity Card

书名/综述主题：
作者：
出版社/期刊：
出版年份：
页数/篇数：
核心主张（一句话）：

§2 — Author and Context [P1 — RECOMMENDED]

• Who is the author? What is their background and credibility?
• Why did they write this? What was the intellectual context?
• How was it received when published?

§3 — Core Arguments / Thesis

• What is the central claim or organizing framework?
• How is the argument structured?
• Key concepts introduced — define precisely [P4]; trigger Concept Decoder if complex

§4 — Chapter-by-Chapter / Paper-by-Paper Synthesis (for Long articles)

• Not a summary of each chapter, but a synthesis of the argumentative arc
• Identify the 3–5 most important ideas

§5 — Strengths [P3]

• What does the work do exceptionally well?
• What new insight does it provide?
• Quality of evidence and argumentation

§6 — Weaknesses and Criticisms [P3 — MANDATORY]

• Logical gaps or unsupported claims
• Missing perspectives or blind spots
• Has the work been criticized in the literature? Cite critics.
• Has subsequent research confirmed or challenged the claims?

§7 — Comparison with Related Works \x3C!-- [I-13 FIX] Changed "2–3 similar works" to "at least 2", consistent with README -->

• Compare with at least 2 similar books/surveys
• What does this work offer that others don't?
• What do the others offer that this one misses?

§8 — Who Should Read This

• Ideal reader profile
• Prerequisites
• How to read it (cover-to-cover? selectively? in what order?)

Citation standard: Full bibliographic entry for the reviewed work; APA format for all other references.

🌍 Category 7: 科普文章 (Science Communication)

Trigger keywords: 科普、面向大众、通俗解释、为什么、怎么理解

Goal: Make a scientific or technical topic accessible to an intelligent non-specialist reader, without sacrificing accuracy or intellectual honesty.

Primary tool: Concept Decoder should be triggered proactively for any concept that would be opaque to a non-specialist. Prefer the "quick" decode mode for science communication.

Required Sections

§1 — The Hook: Why Should You Care?

• Open with a surprising fact, a counterintuitive result, or a relatable scenario
• Connect the topic to something in the reader's daily life
• State the central question in plain language

§2 — The Story: Historical Narrative [P1 — MANDATORY]

• Tell the history as a human story, not a timeline
• Focus on the moments of discovery, confusion, and insight
• Name the people involved (verify [P6]) — science is done by humans

§3 — The Concept: Explained Simply \x3C!-- [I-06 FIX] Added "curious 16-year-old" writing standard, consistent with README -->

• Writing standard: "explain to a curious 16-year-old" — intelligent but without specialist background
• Analogy first, technical term second
• For each technical term introduced: define it immediately in plain language
• Trigger Concept Decoder (quick mode) for any concept requiring more than 2 sentences to explain
• Minimize formulas: if a formula is essential, introduce it as a sentence, not an equation

§4 — The Depth: Going Further

• For readers who want more: one level deeper
• Introduce one or two key technical ideas with careful scaffolding
• This section can be marked as "optional" or "for the curious"

§5 — The Implications: So What?

• What does this mean for science / technology / society?
• What questions does it open up?
• What is still unknown?

§6 — Critical Honesty [P3 — MANDATORY]

• What does the science NOT tell us?
• Where is there genuine uncertainty?
• Common misconceptions to correct
• Avoid hype: distinguish "promising research" from "established fact"

Citation standard: Cite authoritative sources (Nature, Science, major textbooks, institutional reports). Keep citations light in the main text; collect at end as "Further Reading."

Concept Decoder Integration

Concept Decoder (https://clawhub.ai/onlybelter/concept-decoder) is a core dependency of this skill.

Trigger Conditions · 触发条件

When any of the following conditions are met, pause the article draft and call Concept Decoder:

Integration by Category · 各类别集成深度

\x3C!-- [I-07 FIX] Added 7-category integration depth table, consistent with README -->

Embedding Protocol · 嵌入方式

1. Insert a clearly marked block: > 💡 **Concept Spotlight:** [concept name]
2. Call Concept Decoder and embed the Layer 1+2 output (quick) or Layer 1–4 output (standard/deep)
3. Resume article after the spotlight block
4. Cross-reference: "As explained in the Concept Spotlight above, [concept] means..."

Citation Standards

Summary Table · 汇总表

Reference List Format · 参考文献格式示例

\x3C!-- [I-08 FIX] Added three concrete format examples, consistent with README -->

科学论文 / Scientific paper:

[1] Vaswani, A., et al. (2017). Attention is all you need.
    Advances in Neural Information Processing Systems, 30.

书籍 / Book:

[2] Hofstadter, D. R. (1979). Gödel, Escher, Bach: An Eternal Golden Braid.
    Basic Books.

网络资源 / Web resource:

[3] Python Software Foundation. (2024). asyncio — Asynchronous I/O.
    https://docs.python.org/3/library/asyncio.html [Accessed: 2024-01-15]

General Rules · 通用规则

\x3C!-- [I-09 FIX] Added 5th rule about contested claims, consistent with README -->

1. Never cite a source you cannot verify · 不引用无法核实的来源
2. Prefer primary sources over secondary sources · 优先一手来源
3. For scientific claims: cite the original paper, not a blog post about the paper · 科学主张引用原始论文
4. If a claim is uncertain or contested, say so explicitly rather than omitting the citation · 不确定时诚实标注
5. When uncertain about a fact, flag it explicitly: (date unverified), (attribution disputed) · 用标注代替猜测

Error Handling

Topic is too broad · 主题过于宽泛

• Example: /blog 人工智能
• Response: Propose 4–6 sub-topics; ask user to choose one or specify an angle
• Offer a suggested learning path if the user wants a series

Topic is ambiguous · 主题存在歧义

• Example: /blog 熵 (thermodynamic? information-theoretic? philosophical?)
• Response: List the interpretations; ask which angle to take; offer to cover all with clear section separation

Category cannot be inferred · 无法推断类别

\x3C!-- [I-11 FIX] Added A/B/C/D prompt template, consistent with README -->

• Ask the user:

  "这篇文章的目标读者是谁？ A) 专业研究者 / B) 开发者 / C) 大众读者 / D) 学生"

  "Who is the target reader? A) Researchers / B) Developers / C) General public / D) Students"

• The answer usually resolves the category uniquely.

Insufficient information for fact verification · 事实核查信息不足

• Do not fabricate dates, names, or statistics
• Flag uncertain facts explicitly: (date unverified), (attribution disputed)
• Recommend the user verify before publishing
• Suggested verification sources · 推荐核查数据库:
  • 学术文献：arXiv, PubMed, Google Scholar, Semantic Scholar
  • 编程文档：官方文档, GitHub Releases, Changelog
  • 历史事实：Wikipedia（仅作初步核查）+ 原始文献

Topic requires real-time information · 主题需要实时信息

\x3C!-- [I-10 FIX] Added warning template block and recommended database list, consistent with README -->

• Flag clearly with the following warning block:

⚠️ 注意 / Note: 以下内容基于训练数据，可能不包含最新进展。建议发布前在 arXiv / PubMed / 官方文档 / GitHub Releases 中核实。 The following content is based on training data and may not reflect the latest developments. Please verify in arXiv / PubMed / official documentation before publishing.

Formatting Standards

Headers

• H1: Article title only
• H2: Major sections
• H3: Subsections
• Avoid going deeper than H3 in the article body

Math

• Display equations: $$...$$, centered, numbered if referenced
• Inline math: $...$
• Every display equation followed by a plain-language interpretation
• Symbol table for articles with 5+ equations

Code

• Always specify language in fenced code blocks
• Maximum 30 lines per block; break longer examples into labeled parts
• Anti-pattern examples clearly labeled with # ❌ Anti-pattern comment

Lists

\x3C!-- [I-15 FIX] Added nesting depth limit, consistent with README -->

• Use bullet lists for unordered items (features, options)
• Use numbered lists for sequential steps or ranked items
• Do not nest lists deeper than 2 levels

Notes

• This skill generates article drafts; the user should review all factual claims before publishing
• P3 (Critical Thinking) is non-negotiable: a blog post that only presents one side is an advertisement, not an article
• For scientific and mathematical articles, err on the side of more precision rather than less — imprecise popularization is worse than honest complexity
• When in doubt about a fact, say so — intellectual honesty is a feature, not a weakness

Companion Skills · 配套 Skill

\x3C!-- [I-16 FIX] Added structured Companion Skills table, consistent with README -->